Macromolecules 1988,21, 1059-1062 Fr.) 1985,46, 156) and for copolymers by Cantor (Cantor, R. Macromolecules 1981, 14, 1186). (20) de Gennes, P.-G., unpublished results. (21) de Gennes, P.-G. J. Phys. Lett. 1975,36, L.55; J. Phys. Lett. 1978, 49, L.299. (22) de Gennes, P.-G. Coll8ge de
France lectures, 1984.
1059
(23) Pincus, P.; Witten, T . A. Macromolecules 1986, 19, 2609. Pincus, P.; Witten, T . A.; Cates, M. Europhys. Lett. 1986,2, 137. (24) Goldstein, R. J . Chem. Phys. 1986, 84, 3367. (25) Semenov, A. N. Sou. Phys.-JETP (Engl. Transl.) 1985, 61, 733; Zh.Eksp. Teor. Fiz. 1985,88, 1242.
Adsorption of Polybutadienes with Polar Group Terminations on the Solid Surface. 1. Infrared Study at the Silica Surface Masami Kawaguchi,*tMinoru Kawarabayashi,tNobuo Nagata,f Tadaya Kate,+ Akira Yoshioka,fand Akira Takahashit Department of Industrial Chemistry, Faculty of Engineering, Mie University, 1515 Kamihama-cho, Tsu, Mie 514, Japan, and Nippon Zeon Co., Ltd., Research and Development Center, 1-2-1 Yako, Kawasaki-ku, Kanagawa 210, Japan. Received May 22, 1987; Revised Manuscript Received October 21, 1987
ABSTRACT Adsorption of polybutadiene terminated with a very polar functional group (T-PBR)on a silica surface from carbon tetrachloridesolution has been investigated by using IR spectroscopy and compared with normal polybutadienes (PBR). Carbon tetrachloride is a good solvent for polybutadiene. The adsorbed amount, A , the surface coverage 0 of the silanol groups, and the bound fraction of a polymer chain p were measured. A marked difference in adsorption behavior between T-PBR and PBR polymers was observed. T-PBR has an A and 0 twice as large as those for PBR, while the p value was approximately the same for both polymers. From the chemical structureof the terminal group we can qualitatively interpret the difference in the adsorption behavior by taking into account the preferential adsorption of the terminal polar functional groups over the double-bond groups in polybutadiene chains onto the silanol groups of silica surface.
Introduction The loop-train-tail conformation is generally accepted as a structure of flexible polymer chains adsorbed on flat, cylindrical, and spherical surfaces if there is no specific interaction between polymer and surface.l+ Tail portions of adsorbed polymer chain are considered to be randomflight chains attached on a surface with one end, and they play an important role in understanding the behavior of flexible polymer chains a t surfaces and interfaces. In particular, the large layer thickness of the adsorbed polymer chains is believed to be mainly governed by the tail portions in both theoretical and experimental sides. Moreover, the tail portions of adsorbed chains not only act as a steric stabilization moiety for colloidal particles but also can serve as bridges between different particles in flocculation of colloidal dispersion^.^ Tail formation on the solid surface in adsorption experiments is available by use of (1)a polymer terminated with a single specific functional group, which strongly interacts with surface sites of the adsorbing surface, or (2) a block copolymer in which one block adsorbs in train conformation and the other block adsorbs negligibly. The two methods described above contain some demerits: (1) A polymer chain adsorbed with one anchored terminal group may desorb from the surface under good solvent conditions as the result of repulsive interaction of dangling tails caused by the excluded volume effect if the bond strength between the terminal group and a surface site is not very strong. (2) One adsorbed block may not behave as one long train part without loop formation. However, there are a few successful examples. Hadziioannou et recently have measured surface forces between 2-vinylpyridinestyrene block copolymers adsorbed on mica surfaces; the 2-vinylpyridine block adsorbs
*
Mie University. Nippon Zeon Co., Ltd.
strongly on a mica surface with a flattened conformation, while the styrene block is not adsorbed directly to a mica surface a t all. Some block copolymers were used as polymer monolayers at the air-water interface where one block (uncharged block) lays flat on the interface and the other block (charged block) is extended into ~ a t e r . ~ , ~ On the other hand, there are some adsorption studies of polymers terminated by a polar group, which adsorbs preferentially over functional groups in the main polymer chains."'l They used low molecular weight fractions, i.e., M, 2 X lo3. The effect of terminal functional groups on adsorption was definitely observed as expected. For high molecular weights of polymers with terminal groups, a definite effect of terminal functional groups on adsorption behavior has not been recognized. In this study we investigate the effect of a terminal functional group with strong adsorbability on polymer adsorption behavior by using well-characterized samples for both adsorbate and adsorbent. Adsorption measurements of high molecular weight ( lo5)polybutadienes terminated with bis(p-diethy1amino)phenyl)methanol onto the silica surface were performed by using IR spectroscopy in carbon tetrachloride, which is a good solvent for polybutadiene, in comparison with adsorption of normal polybutadiene. The reason why these adsorption experiments are done under good solvent conditions is as follows: under good solvent conditions both the adsorbed amount and surface coverage are usually much smaller than those under poor solvent conditions a t the same polymer concentration; thus,it is expected that the differences in adsorbed amount as well as surface coverage between polymers terminated and those not terminated with functional groups will be emphasized by the preferential adsorption of terminal groups.
-
-
Experimental Section Materials. Polybutadiene (PBR) and polybutadieneterminated with bis(p-(diethy1amino)phenyl)methanol(T-PBR)were kindly supplied from Nippon Zeon Co., Ltd. The chemical
0024-9297/88/2221-lO59$01.50/00 1988 American Chemical Society
Macromolecules, Vol. 21, No. 4, 1988
1060 Kawaguchi et al.
Figure 1. Chemical structure of T-PBR. Table I Molecular Characteristics of T-PBR and PBR contents, % Mn x sample low4 M J M , cis-1,4 trans-1,4 1,2 additive T-PBR-1 36.0 20.2 70.5 1.88 9.3 18.9 71.2 9.9 T-PBR-2 26.0 1.4, 19.1 70.2 1.23 10.7 T-PBR-3 13.0 19.3 70.7 2.56 10.0 PBR-1 36.0 10.0 18.7 PBR-2 21.0 71.3 1.5, structure of T-PBR is shown in Figure 1. PBR was prepared by polymerizing 1,3-butadiene in hexane or benzene with a nbutyllithium catalyst in combination with diglyme (diethylene glycol dimethyl ether) at 40 "C. For T-PBR, after the conversion of polymerization reached 100% a modifier such as 4,4'-bis(diethy1amino)benzophenonewas added to carry out the chemical modification reaction. Supplied polymers were purified by precipitation from their benzene solutions into a large amount of methanol by stepwise addition. GPC measurements of PBR and T-PBR samples prepared in the same batch give similar chromatograms, and this result indicates that T-PBR series make no complex via hydrogen bonding of the polar terminal groups in solution. After the precipitated samples were dried under vacuum they were dissolved again in benzene and fractionated into more than 10 fractions by addition of methanol at 25 "C. The fractionated samples were freeze-dried from benzene solution. The samples so prepared were stored in a brown glass desiccator for protection from light; the desiccator was filled with Ar gas to avoid oxidation. The number average molecular weights of the fractionated polymers were determined by using a Hewlett-Packard 502 high-speed membrane osmometer in toluene or carbon tetrachloride at 25 " C . A Sartorius SM 67 N 11539 filter was used as a semipermeable membrane. We chose two fractionated PBR samples and three fractionated T-PBR samples for the adsorption studies. The polydispersities of both T-PBR and PBR samples were determined from GPC measurements by using polystyrene samples with narrow molecular weight distributions as a standard. The microstructures such as cis-1,4, trans-l,4, and 1,2 additive of PBR and T-PBR samples were determined from 13C NMR spectra of 7% solution in deuteriwhloroformat room temperature by using a Nippon Denshi (Model FX-100) spectrometer.12 The molecular characteristics of the polymer used in this study are listed in Table I. The nonporous Aerosil 130 silica (Degussa A. G., West Germany) was used as the adsorbent after being heated at 120 " C . According to the manufacturer, the average particle diameter is 160 A with a surface area of 141 m2/g, and the surface concentration of silanol groups on the silica surface is 3 silanol groups/lOO
At. Carbon tetrachloride used as a solvent for polymer adsorption experiments is purified by a fractional distillation just before use. Polymer Adsorption. A 20-mL polymer solution with known concentration was mixed with 0.16 g of silica in a stoppered glass tube for centrifugation. After equilibration by sitrring with a magnetic stirrer chip at 25 "C the supernatant solution was separated to sediment the silica particles by using a Kubota KB-200 B centrifuge. The equilibrium concentration C in the supernatant solution was determined by using a JASCO &/IR-3 IR spectrometer at a characteristic wavenumber of 2920 cm-', which corresponds to the streching vibration of CH groups. The sedimented silica gel was placed in an IR cell made from 4 X 4 cm2KBr plates, and it was subjected to the IR spectrometer with the reference cell containing carbon tetrachloride or the supematant solution. To confirm the reproducibilityof adsorption measurements we performed at least two measurements for the same concentration. The errors in the adsorbed amount are less than 5% whereas those in both B and p are less than 10%.
I
I
I
0.1
0.2
0.3
cpx102
/9"'
-
Figure 2. Adsorption isotherms of T-PBR and PBR polymers: (0)T-PBR-1; (8)T-PBR-2; ( 0 )T-PBR-3; (0) PBR-1; (B)PBR-2.
Wave Number /cm'
Figure 3. IR spectra of Aerosill30 silica in carbon tetrachloride: A, silica; B, silica on which PBR-1 adsorbs at a dosing concentration Co of 0.2 g/100 mL (equilibrium concentration in the supernatant solution C, = 0.164 g/100 mL); C, silica on which T-PBR-2 adsorbs at Co of 0.1 g/100 mL ( C , = 0.046 g/100 mL). Table I1 Adsorution Data at the Plateau Region of T-PBR and PBR A
sample T-PBR-1 T-PBR-2 T-PBR-3 PBR-1 PBR-2
X
lo8,
g cm-2
0
6.1
0.31
6.2
0.34 0.25
6.3 3.2 3.3
0.18 0.13
P 0.14
0.15 0.11 0.15 0.11
Results T h e cis-1,4, trans-1,4, a n d 1,2 additive contents of all polymer samples used here are almost the same, as seen from Table I. In Figure 2 all adsorption isotherms of PBR and T-PBR are illustrated. T h e adsorbed amount A , expressed in g/cm2, increases with t h e equilibrium concentration C, and reaches a plateau value at higher C,. T h e shape of all adsorption isotherms is t h e rounded curve, reflecting the somewhat wide molecular weight distribution of T-PBR a n d PBR samples a n d not their high-affinity character. Adsorption isotherm curves are independent of molecular weight for t h e respective series. W e notice that the plateau adsorbed a m o u n t of T-PBR is twice as large as t h a t of P B R , as seen in Table 11. Figure 3 shows IR spectra of the silica gel: A, the silica suspended in carbon tetrachloride; B, t h e silica equilibrated with the plateau adsorbed amount of PBR; a n d C, the silica equilibrated with the adsorbed amount of T-PBR one lower t h a n the plateau value. T h e absorption peak of silanol groups in carbon tetrachloride was observed at 3690 cm-'. T h e r e are two characteristic peaks in the IR spectra of B a n d C: one peak at 3690 cm-' corresponds to the absorbance of free silanol groups a n d another peak at -3100 cm-' is d u e t o t h e absorbance of CH streching vibration of adsorbed polymer chains. T h e absorbance at -3100 cm-' in Figure 3C is stronger t h a n that in Figure 3B since the magnitude of adsorbed amount of T-PBR (C)
Adsorption of Polybutadienes on Si 1061
Macromolecules, Vol. 21, No. 4, 1988
a
02
p-01
CpX102
0.2
0.3
OO
12. 01
g.0."
02
1
0.3 1
cpx102
mml
Figure 5. Plots of p vs the equilibrium concentration Cp in the supernatant solution. Symbols are the same as in Figure 2.
/9"
Figure 4. Plots of 8 vs the equilibrium concentration C , in the supernatant solution. Symbols are the same as in Figure 2.
is larger than that of PBR (B). In general, when the polymer chains adsorb on the silanol groups via an interaction such as hydrogen bonding the frequency of silanol groups should shift lower; that is, the frequency shift of vibrations of silanol groups is observed to be due to the perturbed silanol groups. In this study the particular frequency shift of silanol groups was not apparent in IR spectra of the silica on which the polymers adsorb. However, the presence of the peak at -3100 cm-' in the IR spectra of B and C indicates that both T-PBR and PBR molecules adsorb on the silica surfaces. Moreover, the concentration of free silanol groups was actually changed to decrease after adsorption of TPBR or PBR. These facts mean that both polymers interact with the silanol groups. The previous works showed that the double bonds in polybutadiene13J4and some organic compound^'^'^ containing the CH=CH group induce an IR frequency shift of the perturbed silanol groups when they adsorb on the silica surface. The reason for the weak interaction between double bonds and silanol groups in this study is not clear, but it may be due to the much higher content of the 1,2 additive, Le., the vinyl bond, in T-PBR and PBR samples in comparison with polybutadienes used in our previous study.13 The fraction 0 of the silanol groups occupied by adsorbed polymer chains can be calculated from 0 = (C, - CO€I)/C,
a Q 2 1 e m
(1)
where C, and COHare the numbers of silanol groups before and after adsorption, respectively. Plots of 6' vs the equilibrium concentration C, in the supernatant solution are displayed for all polymers in Figure 4. The 6' value increases with C, and reaches a constant value above C, = 0.1 g/100 mL. The plateau value 0 shows a slight molecular weight dependence; that is, high molecular weight samples have a somewhat larger 0 than low molecular weight ones for the respective polymers. For PBR series the higher molecular weight species tend to have higher 8; however, for T-PBR series the similar tendency was not observed. The difference in 0 between T-PBR-1 and TPBR-2 exceeds experimental errors. In this stage we have no explanation for this difference. The plateau values of 0 tabulated in Table I1 show that their magnitudes for T-PBR are almost twice as large as those for PBR. We can calculate the fraction p of monomer units attached directly to the silanol groups by the following equation, assuming that a double bond in butadiene chains should interact with one silanol group: p = (weight of monomer units directly attached to silanol groups per unit area)/(weight of adsorbed amount per unit area) (2) The p values so determined are plotted against C, in Figure 5. These curves are almost independent of molecular weight as well as polymer species because of the
twice larger adsorbed amount and 0 of T-PBR than of PBR. The plateau value of p listed in Table I1 is taken as -0.1.
Discussion We can notice some features from the plateau values of adsorption data listed in Table 11. The plateau values of adsorbed amount as well as p for PBR are lower in comparison with most data reported p r e v i o u ~ l y . ~This ~ ' ~ fact is understandable since polybutadiene is a less polar polymer and has complicated structures involving three different microstructures and also since carbon tetrachloride is a very good solvent for polybutadiene, showing a large second virial coefficient of -2.0 X mol/g2. Consequently, the relative strength of the bond between double bonds of butadiene and silanol group in comparison with the bond between solvent and silanol groups is weaker than for the other polymer-solvent-silica combinations, and thus no evident frequency shift of the occupied silanol groups is observed. Another feature point is the twice larger adsorbed amount and 6' for T-PBR than for PBR. The terminal group of T-PBR contains some specific functional groups such as hydroxyl, amine, and phenyl as seen from its chemical structure (see Figure 1). Adsorption experiments of organic compounds containing such functional groups on the silica surface in carbon tetrachloride by using IR spectrometer showed that amines and alcohols form tight hydrogen bonds with silanol groups, respectively, and thus shift the perturbed silanol groups to a lower frequency as compared with the frequency of unperturbed silanol groups.16J8-20 In this experiment the absence of an evident frequency shift of silanol groups perturbed by the terminal groups of T-PBR is due to the extremely lower number of the terminal groups compared to butadiene monomeric units. The evidence for the stronger interaction between the terminal group of T-PBR and silanol groups is proved by the displacement adsorption of T-PBR and PBR from the silica surface by addition of displacers such as benzene, dioxane, and acetone.21 The displacement adsorption by adding 1mL of benzene showed that T-PBR chains are not desorbed from the silica surface, while PBR chains are completely exchanged with displacer molecules. However, the addition of 1 mL of dioxane or acetone leads to a complete desorption of both T-PBR and PBR chains from the silica surface. Therefore, three functional groups in the terminal group of T-PBR chain more strongly interact with the silanol groups than double bonds in polybutadiene chains. We now propose a rather speculative model with which we wish to explain a t least qualitatively our observations. The terminal groups preferentially adsorb on the silanol group over the double bonds in polybutadiene chains a t an earlier adsorption time. In this stage the terminal group only interacts with the silica surface, while the remaining part, i.e., a polybutadiene chain, behaves as a long tail. However, this preferential adsorption of the terminal groups of T-PBR chains will be reduced with increasing
1062
Macromolecules 1988, 21, 1062-1070
adsorption time because of blocking of the unoccupied silanol groups by earlier arrivals, in which the overall chain conformation of one end-attached T-PBR chain is not much different from the free chain conformation suggested by Hesselink22and Tanaka.23 Subsequently, the double bonds in butadiene parts of T-PBR chains adsorb on the silanol groups in order to attain an equilibrium state, and finally an adsorbed amount twice as large as that of PBR chains is established for T-PBR chains. The bound fraction p is sometimes useful to infer the conformation of adsorbed polymer chain. Since the plateau value of p is nearly constant irrespective of polymers, one can imagine that the conformation of adsorbed T-PBR chains is not different from that of adsorbed PBR chains. However, for T-PBR chains their one end is undoubtedly attached to the silica surface due to the high adsorbability of the terminal group. The polymer segments located close to the anchored end adsorb easily on the silica surface since their mobility should be restrained in comparison with that of unadsorbed segments located far from the adsorbed end segment. Therefore, in the equilibrium state T-PBR chains take a conformation consisting of one relative long tail and some short loops as well as trains. On the other hand, the adsorbed PBR chains will take the generally accepted loop-train-tail conformation, where both ends are dangling. The difference between both adsorbed conformations may be explored by the measurement of adsorbed layer thickness. A preliminary ellipsometric study24 of adsorption of T-PBR and PBR onto a platinum plate from carbon tetrachloride solution showed that the thickness of T-PBR is thicker than that of PBR and the adsorbance of T-PBR is -1.5 times that of PBR. These ellipsometric data seem to be convenient for more adsorbed polymer on the surface for the T-PBR chains. The plateau data on p and 6 for T-PBR imply on the order of 300 covered adsorbed segments per chain ( p 0.1) and on the order of 600 silanol groups per particle; Le., two molecules are adsorbed on the silica surface if no bridging between different silica particles is assumed. T-PBR chains whose radii of gyration is on the order of 200 A will be able to wrap around the silica particle well by taking the suggested structure of adsorbed chains.
-
Conclusions High molecular weight polybutadienes terminated with a single very polar group show a different adsorption behavior a t the silica surface in carbon tetrachloride as compared with normal polybutadiene. The former has an adsorbed amount and a surface coverage of the silanol groups that are twice as large as the same features in the
latter, but the bond fraction p (i.e., the fraction of monomers of adsorbed chains being in contact with the surface) is almost equal for both polymers. This difference is rather speculatively interpreted by taking into account the preferential adsorption of terminal polar groups to the silanol groups over the double bonds in butadiene chains. We wish to emphasize that a chain end terminated with polar groups is very important for the determination of adsorption behavior of polymer chains. Registry No. SiOz, 7631-86-9. References a n d Notes Vincent, B.; Whittington, S. In Surface and Colloid Science; Matijevic, E., Ed.; Plenum: New York, 1981;p 1. Tadros, Th. F. In The Effect of Polymers on Dispersion Properties; Tadros, Th. F., Ed.; Academic: London, 1982;p 1. Takahashi, A.; Kawaguchi, M. Adu. Polym. Sci. 1982,46, 1. Fleer, G. J.; Lyklema, J. In Adsorption from Solution at the SolidlLiquid Interface; Parfitt, G. D., Rochester, C . H., Eds.; Academic: London, 1983;p 153. Napper, D. H.Polymeric Stabilization of Colloidal Dispersions; Academic: London, 1983. Hadziioannou, G.; Patel, S.; Granick, S.; Tirrell, M. J . Am. Chem. SOC.1986,108,2869. Kawaguchi, M.;Oohira, M.; Tajima, M.; Takahashi, A. Polym. J. 1980,12, 849. Bringuier, E.;Vilanove, R.; Gattot, Y.; Selb, J.; Rondelez, F. J. Colloid Interface Sci. 1985,104, 95. Crowl, V. T.; Matati, M. A. Discuss. Faraday SOC. 1966,42, 301. Crowl, V. T. J. Oil Colour Chem. Assoc. 1967 50,1023. Sakai, I.;Nakamae, K.; Ohkawa,T.;Inubushi, H.; Mastumoto, T. Kobunshi Ronbunshu 1977,34, 661. Yoshioka, A.; Komuro, K.; Ueda, A.; Watanabe, H.; Akita, S.; Masuda, T.; Nakajima, A. Pure Appl. Chem. 1986,53, 1697. Kawaguchi, M.; Sano, T.; Takahashi, A. Polym. J. 1981,13, 1019. Kawaguchi, M.;Aoki, M.; Takahashi, A. Macromolecules 1983, 16, 635. Little, L. H.; Mathieu, M. V. Actes Congr. Int. Catal., 2nd 1960,771. Marshall, K.; Rochester, C. H. J. Chem. SOC.,Faraday Trans. I 1975,71, 1754. Howard, G. J. In Interfacial Phenomena in Apolar Media; Eicke, H. F., Parfitt, G. D., Eds.; Dekker: New York, 1987;p 281. Griffiths, D. M.; Marshall, K.; Rochester, C. H. J . Chem. SOC., Faraday Trans. 1 1974,70,400. Curthoys, G.; Davydov, V. Ye.; Kiselev, A. V.; Kiselev, S. A.; Kuznetsov, B, V. J. Colloid Interface Sci. 1974,48, 58. Rochester, C. H.Adv. Colloid Interface Sci. 1980,12, 43. Kawaguchi, M.; Kawarabayashi, M.; Takahashi, A., unpublished data. Hesselink, F. Th. J. Colloid Interface Sci. 1975,50, 606. Tanaka, T. Macromolecules 1977,10,51. Kawaguchi, M.; Kawarabayashi, M.; Takahashi, A., unpublished data.
Molecular Rheology of H-Polymers T.C . B. McLeish Cavendish Laboratory, Madingley Road, Cambridge CB3 OHE, United Kingdom. Received July 8, 1987
ABSTRACT The rheological behavior of branched polymer melts contrasted with that of linear melts motivates a discussion of a molecular model of a melt of H-polymers. Concentrating on the dynamics of the "backbone", . ~ ~ model we derive and discuss the linear viscoelastic properties in the light of experiments by R o o v e r ~ The is extended to treat nonlinear deformations, with path-length extension treated in a self-consistent way, and compared with rheological behavior of branched LDPE.
Introduction The primitive path or tube model of polymer melts as developed by Doi and has met with consider-
able success when applied to the dynamics of linear polymers."' The extension of the theory to star polymers has been mades,g,zoand is in accord with both numericallo
0024-9297/88/2221-1062$01.50/0 0 1988 American Chemical Society
